Map of Content - Knowledge Architecture
Navigational index for notes about how to design knowledge systems, AI memory, context engineering, and the emerging Agent Web.
The Fundamental Problem
- 2026-04-10-context-fragmentation-ai-memory - AI memory is siloed within platforms; the primary obstacle to AI productivity is lack of persistent shared context, not model intelligence
The Two Webs
- 2026-04-10-agent-web-vs-human-web - Information architecture splits into Human Web (UI, folders, for reading) and Agent Web (APIs, embeddings, for machines); future-proof knowledge must work for both
Building the Solution
- 2026-04-10-open-brain-tech-stack - Platform-agnostic personal memory: Postgres (ownership) + PGVector (semantic search) + MCP (universal interface); prevents lock-in
- 2026-04-10-context-engineering-advantage - Pre-loading AI with accumulated roles, projects, and constraints before prompting; the compounding productivity gap between chatbot users and “Open Brain” users
Depth Over Volume
- 2026-04-10-concentrated-depth-compounds-over-scattered-effort - Concentrated context investment compounds; starting from zero every session dissipates effort - appears in both AI workflows and devotional practice
The Human Benefit
- 2026-04-10-machine-readability-for-thought-clarity - Designing for machine readability forces precision in human communication; reduces ambiguity; the “person with least context” forcing function
Institutional Risk
- 2026-04-10-ai-architectural-lockin-is-process-based - Lock-in is workflow and process, not subscription cost; AI tool selection is an architectural commitment